fineui-window

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from iframe components and user input dialogs.
  • Ingestion points: Data is received via the closeArgument parameter from child windows and input strings from Prompt dialogs.
  • Boundary markers: No explicit boundaries are defined for the transferred data.
  • Capability inventory: The skill includes methods to update parent UI controls and trigger server-side events based on received arguments.
  • Sanitization: Documentation explicitly warns against using RawHtml for user-supplied input to prevent injection.
  • [DYNAMIC_EXECUTION]: The skill documents APIs for cross-window script execution.
  • Evidence: The hideExecuteScript and GetHideExecuteScriptReference methods allow a child iframe to run JavaScript in the parent window.
  • Execution Method: Scripts are passed as strings and executed at runtime by the framework.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 12:55 PM